Revenue rarely disappears in one dramatic loss. It drains through seven ordinary seams in the lifecycle, each individually forgivable and collectively expensive. Here is where to look and how to quantify each one from data you already have.
Ask a revenue leader where the quarter went and you will usually get a story about two or three deals. Those deals are real, but they are rarely where the money went. The money went out through seams — small, structural, unglamorous gaps in the lifecycle that leak a little on every opportunity rather than a lot on a few.
Leakage of that shape has a signature: it survives. Change the rep, change the manager, change the comp plan, and the same percentage still disappears, because nothing about the system that produced it changed. That is what makes it worth finding — a structural leak, once closed, stays closed.
1. Inbound interest that ages before anyone touches it
The first leak happens before a deal exists. A lead arrives, sits in a queue, gets routed to someone on holiday, and is worked forty hours later. Nothing about the lead changed in those forty hours except the buyer's attention, which moved on.
Measure it as the distribution, not the average. The mean response time flatters you because the fast ones dominate the count; what costs money is the tail. Segment by source and by hour of arrival, and the pattern usually resolves into something dull and fixable: nobody owns Friday afternoon, or a routing rule sends a whole segment to a queue with one person in it.
2. Opportunities in the wrong stage
Stages are supposed to describe what has happened. In practice they describe what the rep is comfortable reporting. A deal with no economic buyer identified and no mutual plan sits in Negotiation because it was there last month and moving it back is a conversation nobody wants.
The diagnostic is behavioural rather than declared: compare each deal's activity signature against deals that genuinely reached that stage and closed. Meeting cadence, number of contacts engaged, whether procurement has appeared, whether pricing has been discussed in writing. Deals whose behaviour does not match their stage are mispriced in the forecast, and the error compounds because stage drives every downstream weighting.
3. Silent pipeline
Every pipeline contains deals that nobody has touched in weeks and nobody has closed. They are not lost — losing them would require a decision. They simply sit, inflating coverage and depressing the conversion rates of every cohort they belong to.
The threshold that matters is not a fixed number of days but the median gap for deals of that size and stage that eventually closed. A fourteen-day silence in a transactional segment is a different fact from a fourteen-day silence on an enterprise deal in legal review.
A deal nobody has touched in a month is not in your pipeline. It is in your spreadsheet.
The distinction most coverage ratios ignore
4. Slippage that is never priced in
Close dates move. That is not the leak. The leak is that they move in one direction, predictably, by an amount the organisation could estimate but does not — so every forecast is built on dates that a historical base rate would have already corrected.
Compute slippage per segment as the distribution of (actual close date − first-committed close date). If a segment slips a median of eleven days, a forecast that takes rep-entered dates at face value is systematically early, every quarter, by roughly eleven days' worth of revenue at the boundary.
5. Expansion that nobody was watching for
The most expensive leaks are usually the ones that never appear in a report, because a report shows you what happened and an expansion that nobody noticed did not happen.
The signal is comparative: accounts that resemble your best expansions on the dimensions that mattered — product usage shape, team growth, support engagement, champion seniority — but where no second product was ever proposed. This is the leak that is invisible to every pipeline review, because it is not in the pipeline.
6. Renewals decided long before the renewal conversation
By the time a renewal appears on a calendar, the decision has usually been made. The information that would have predicted it was scattered across support tickets, usage decline, a champion who changed jobs, and a sentiment shift in the last two QBRs — none of which are in the CRM, and all of which are in systems you already pay for.
Joining that history to the renewal calendar is not a modelling problem. It is a plumbing problem that most organisations never get to because it spans three tool owners.
7. The handoffs
Marketing to sales. SDR to AE. AE to onboarding. Onboarding to CS. Each handoff loses context, and the loss is asymmetric: the receiving side does not know what it was not told, so it never asks.
The measurable version is time-in-transition and re-discovery cost — how long an account sits between owners, and how much of the first call after a handoff is spent re-establishing things the previous owner already knew.
How to quantify before you fix
The ordering matters more than the list. Every leak above is real in most organisations, and fixing the wrong one first is how a programme loses its sponsor. Quantify each in currency, with a stated method and a stated confidence, then fix in descending order of recoverable value.
- Define the leak as a measurable event, not a feeling — 'opportunities with no activity for more than the segment median' beats 'stalled deals'.
- Count the population and attach value at the deal's own amount, not an average.
- Apply a recovery rate you can defend from your own history, not a vendor's benchmark.
- State the confidence interval. A range you believe is more useful than a point estimate nobody does.
- Re-measure after the fix. A leak that was closed and re-opened is the most common failure mode, and only measurement catches it.
The pattern underneath all seven
None of these leaks are caused by people being bad at their jobs. They are caused by feedback loops that are too slow: the information that would have prevented the loss existed, but it arrived after the moment it mattered, in a system nobody was looking at, in a form nobody could act on.
That is why adding another dashboard rarely helps. A dashboard shortens the distance between data and a human who has to notice it. What actually closes a structural leak is shortening the distance between the signal and the action — which means detection, evidence and a prepared next step arriving together, at the moment the signal fires.